Computer Vision
Kitware’s Computer Vision team is a leader in the creation and support of state-of-the-art software technology. We provide robust solutions to academic and government institutions, and commercial corporations worldwide. Our technology centers on automated image and video analysis and analytics, which are applicable to a variety of fields from security and defense to business intelligence. We have deployed operational systems, including a wide-area motion imagery (WAMI) tracker tracking system transitioned for Intelligence, Surveillance, and Reconnaissance (ISR) in to theatre and content-based image retrieval capabilities being transitioned tothrough DARPA and AFRL programs.
What Kitware Can Do For You
  • Develop state-of-the-art algorithms for mover detection, tracking, and classification.
  • Build event-based video content extraction capabilities to filter and prioritize archived or streaming video data.
  • Refine and integrate academic research into operational prototypes running on operational data.
  • Build advanced visualization and GUI tools enabling analysts to intuitively examine extracted video content.
  • Develop end-to-end solutions for the most challenging problems in automated image and video analysis.
Featured Video: FMV & WAMI Detection & Tracking
Kitware has developed operational, state-of-the-art tools and robust algorithms for target detection, tracking, and classification in wide-area motion imagery (WAMI) and full motion video (FMV).

   
Case Studies
Content-based Image & Video Retrieval
Customer challenge: DARPA and AFRL programs needed to efficiently search the content of FMV and WAMI video.
Kitware Solution: We developed a robust system for content-based image retrieval. The system ingests, pre-processes, and stabilizes an incoming video feed and then identifies and characterizes moving objects, both dismounts and vehicles. More complex video descriptors, such as motion descriptors, (including kinematic, deformable, and articulated motion), appearance descriptors (including color and shape), and behavior descriptors (such as running, carrying, vehicle u-turn and many others) characterize deeper semantic content. This semantic analysis enables live alerting to special ops personnel and can cue unexpected or suspicious activity in a video stream, which can be critical for mission success. Further, advanced forensic capabilities allow retroactive retrieval of activities of interest from large datasets.
Social Media Understanding for Intelligence
Customer challenge: Need to effectively search large, crowd-source image and video collections for relevant content for intelligence purposes.
Kitware solution: We developed a web-enabled multimedia search algorithm that characterizes content based on pixel and audio information and creates efficient search indicies. The semantic, keyword-based search results can be iteratively refined by the user to train the system. We have tested the system on 150,000 videos that total more than 1.2K hours of content. This enables analysts to rapidly search, interactively refine, and discover videos of interest from a large-scale video archive to create actionable intelligence.
Multiple Target Tracking in FMV and WAMI
Customer need: Robust system to identify and track moving targets in full motion video (FMV) and wide area motion imagery (WAMI)
Kitware Solution: We developed and deployed in theatre an operational system that automatically detects and tracks multiple targets in low resolution and low frame data from FMV, WAMI, and low-quality ground cameras. It enables the customers to exploit real-time processing software for small, difficult-to-detect and high-value target tracking across challenging conditions. We have started to put this tracker, in addition to our other image and video analysis capabilities, on Forge.mil as part of the Kitware Image and Video Exploitation and Retrieval Toolkit (KWIVER), with unlimited rights to the government.